Skip to main content

The Rise of NVIDIA, Part 4: Inventing the GPU

Inventing the GPU

🎬 Prefer to watch? Here's the 60-second version:

Watch on YouTube ▶

🎮 The Rise of NVIDIA — a 20-part series. See all parts »  |  « Part 3: Thirty Days from Death

By the start of 1999, Nvidia was no longer the company that had nearly died. The RIVA TNT had proven it could ship a competitive 3D chip on a brutal cadence, and the money was finally flowing in the right direction. But Jensen Huang wanted more than another fast graphics card. He wanted to redraw the map of the industry so that Nvidia sat at the center of it — and in a single year, with one chip and one carefully chosen word, that is exactly what happened.

Going public at twelve dollars

The first move was financial. On January 22, 1999, Nvidia listed on the Nasdaq under the ticker NVDA, pricing its initial public offering at $12 per share. It was a modest debut by the standards of the dot-com frenzy swirling around it — Nvidia sold chips, not clicks, and the market was busy chasing companies with no revenue and enormous promises. A graphics hardware firm from Santa Clara was not the headline of the day.

Yet the timing was deliberate. An IPO gave Nvidia a war chest and a currency — public stock — to recruit engineers and outspend rivals on research. It also imposed discipline: quarterly scrutiny for a company that already ran on a self-imposed six-month product cycle. Founded in 1993 by Huang, Chris Malachowsky, and Curtis Priem, Nvidia was now answerable to the market. The $12 share would become one of the most famous entry points in financial history, but in early 1999 it simply bought the company room to swing for something bigger.

The word that made a category

On August 31, 1999, Nvidia announced the GeForce 256 and did something no chip company had done before: it gave its product a brand-new species name. The GeForce 256, Nvidia declared, was “the world’s first GPU” — a Graphics Processing Unit. And crucially, Nvidia did not leave the term vague. It defined a GPU precisely as “a single-chip processor with integrated transform, lighting, triangle setup/clipping, and rendering engines that is capable of processing a minimum of 10 million polygons per second.”

The genius of that sentence was that Nvidia wrote the definition and then shipped the only product that met it. Competitors making 3D accelerators were suddenly, by Nvidia’s own yardstick, selling something lesser. The name “GeForce” itself fused “geometry” with “force,” a nod to the geometry engine at the heart of the chip. The “256” came from its 256-bit QuadPipe rendering engine — four 64-bit pixel pipelines working in parallel. It was marketing and engineering welded together, and it worked. Twenty-six years later, we still call these chips GPUs. Nvidia didn’t just win a product cycle; it named the field it intended to dominate.

What actually changed inside the silicon

Behind the branding sat a genuine architectural leap. Earlier 3D accelerators handled rasterization — turning geometry into pixels — but leaned on the computer’s main CPU to do transform and lighting, the math that positions every vertex in 3D space and calculates how light falls on it. That work was heavy, and it throttled how many polygons a scene could contain.

The GeForce 256, codenamed NV10 and built by TSMC on a 220-nanometer process, pulled that math onto the graphics chip itself with a dedicated hardware transform and lighting (T&L) engine. Offloading it freed the CPU and let games render richer, more detailed worlds. The chip also added hardware motion compensation for MPEG-2 video, cube environment mapping, and dot-product bump mapping — features that made surfaces look convincingly three-dimensional. It was the first fully Direct3D 7-compliant accelerator, and it shipped on October 11, 1999, with a faster DDR-memory edition following on December 13.

There was a quieter consequence, too. Hardware T&L had until then lived in expensive workstation cards aimed at computer-aided design. By bringing it to a consumer part, Nvidia collapsed the cost of that capability — and then turned around and used the same NV10 silicon to launch Quadro, its own professional line for the CAD market. One chip, two markets, and a template Nvidia would reuse for decades: build the best core, then aim it everywhere at once.

In twelve months, Nvidia had raised its own capital, coined the vocabulary of an industry, and shipped the product that justified the word. The GeForce 256 was not the fastest thing on earth for long — rivals and Nvidia’s own successors would eclipse it within a year — but it was the moment the company stopped following the market and started defining it.

The stakes were not abstract. The late-1990s 3D market was fiercely contested, with 3dfx’s Voodoo cards enjoying near-cult status among PC gamers and rivals like ATI and Matrox pushing hard. By fusing a geometry engine, hardware lighting, and four pixel pipelines into one part that no competitor could yet match on its own terms, Nvidia didn’t just release a faster card — it changed what “fast” meant. The companies that had defined 3D gaming a year earlier now had to catch up to a spec someone else had written.

Next in the series: the year 2000, when Nvidia absorbed a fallen giant and turned an old enemy’s crown jewels into its own.


🔗 Explore more from Syncster

Comments

Popular posts from this blog

Cursor AI Review: Is the AI Code Editor Worth It?

I've been using Cursor as my main code editor for a while now, and enough people have asked whether it's worth switching to that a proper review felt overdue. Short version: for me, yes — but with caveats. What is Cursor? Cursor is an AI-first code editor built as a fork of VS Code. That means every extension, theme, and keybinding you already use in VS Code works here, but with AI woven directly into the editing experience instead of bolted on as a plugin. It's made by Anysphere and can run models from OpenAI and Anthropic under the hood. What I like Tab completion is uncanny. Cursor predicts your next edit — not just the rest of the line, but the next change across the file. Once you get used to hitting Tab, going back to a plain editor feels slow. The Composer / Agent mode. You describe a change in plain language and it edits multiple files at once, showing you a diff to accept or reject. For refactors and boilerplate, this saves real time. It unde...

How I used Google Sheets and Apps Script

Google Sheet is one of the most powerful spreadsheet application that exists online, rivaling with Microsoft's Excel. One of the main strengths is its strong support for collaboration with other users, much easier and popular than collaboration tools with Microsoft Office. Aside from plain spreadsheet, it also supports extensions such as macro. If you are familiar with macros on other office tools, they work almost the same. However, the most extension I use and tinker with is the Apps Scipt . Apps Script Extension One of the challenges I faced recently is how do I track or monitor reports in our department if they are submitted on time or worst, forgotten due to lack of better monitoring tools. So I thought if there can be simple applications that can be deployed or use by a more general user to allow reminding periodically what reports are approaching due dates or those that are past dues. Then I looked for a way, instead of creating a full blown app from scratc...

MacBook Pro M5 vs M5 Pro: Which One Should You Actually Buy?

Apple's latest 14-inch MacBook Pro comes in two very different flavors: the base M5 and the step-up M5 Pro . On paper they look similar — same gorgeous Liquid Retina XDR display, same design — but under the hood the gap is bigger than the names suggest. Here's a clear, no-hype breakdown, with concrete use cases so you can match the chip to your work. Quick spec comparison Spec M5 M5 Pro CPU 10-core (4 performance + 6 efficiency) Up to 18-core (6 performance + 12 efficiency) GPU 10-core Up to 20-core Neural Engine 16-core 16-core Memory bandwidth 153 GB/s 307 GB/s (roughly double) Unified memory 16 / 24 / 32 GB 24 / 48 / 64 GB Max storage Up to 4 TB SSD Up to 8 TB SSD Battery (video playback) Up to 24 hours Up to 22 hours Media engines Single encode/ProRes engine More encode/ProRes engines (higher configs) What actually changes between them More cores — the M5 Pro nearly doubles CPU cores and adds GPU cores, so sustained, multi-threaded work finishe...